Mercurial > repos > davidvanzessen > argalaxy_tools

comparison imgt_loader/imgt_loader.py @ 19:672d5e010b1f draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author davidvanzessen
date Fri, 25 Dec 2015 05:35:06 -0500
parents 431797cd74c8
children 2d38d7c1d0d0
comparison
equal deleted inserted replaced
18:431797cd74c8 19:672d5e010b1f
24 old_sequence_columns = [u'CDR1-IMGT', u'CDR2-IMGT', u'CDR3-IMGT'] 24 old_sequence_columns = [u'CDR1-IMGT', u'CDR2-IMGT', u'CDR3-IMGT']
25 old_junction_columns = [u'JUNCTION'] 25 old_junction_columns = [u'JUNCTION']
26 26
27 added_summary_columns = [u'Functionality', u'V-REGION identity %', u'V-REGION identity nt', u'D-REGION reading frame', u'AA JUNCTION', u'Functionality comment', u'Sequence'] 27 added_summary_columns = [u'Functionality', u'V-REGION identity %', u'V-REGION identity nt', u'D-REGION reading frame', u'AA JUNCTION', u'Functionality comment', u'Sequence']
28 added_sequence_columns = [u'FR1-IMGT', u'FR2-IMGT', u'FR3-IMGT', u'CDR3-IMGT', u'JUNCTION', u'J-REGION', u'FR4-IMGT'] 28 added_sequence_columns = [u'FR1-IMGT', u'FR2-IMGT', u'FR3-IMGT', u'CDR3-IMGT', u'JUNCTION', u'J-REGION', u'FR4-IMGT']
29 added_junction_columns = [u"P3'V-nt nb", u'N1-REGION-nt nb', u"P5'D-nt nb", u"P3'D-nt nb", u'N2-REGION-nt nb', u"P5'J-nt nb", u"3'V-REGION trimmed-nt nb", u"5'D-REGION trimmed-nt nb", u"3'D-REGION trimmed-nt nb", u"5'J-REGION trimmed-nt nb"] 29 added_junction_columns = [u"P3'V-nt nb", u'N-REGION-nt nb', u'N1-REGION-nt nb', u"P5'D-nt nb", u"P3'D-nt nb", u'N2-REGION-nt nb', u"P5'J-nt nb", u"3'V-REGION trimmed-nt nb",
30 u"5'D-REGION trimmed-nt nb", u"3'D-REGION trimmed-nt nb", u"5'J-REGION trimmed-nt nb"]
30 31
31 outFile = args.output 32 outFile = args.output
32 33
33 #fSummary = pd.read_csv(triplets[0][0], sep="\t", low_memory=False) 34 #fSummary = pd.read_csv(triplets[0][0], sep="\t", low_memory=False)
34 fSummary = pd.read_csv(args.summ, sep="\t", dtype=object) 35 fSummary = pd.read_csv(args.summ, sep="\t", dtype=object)
61 for col in added_junction_columns: 62 for col in added_junction_columns:
62 tmp[col] = fJunction[col] 63 tmp[col] = fJunction[col]
63 64
64 outFrame = tmp 65 outFrame = tmp
65 66
66 outFrame.columns = [u'ID', u'VDJ Frame', u'Top V Gene', u'Top D Gene', u'Top J Gene', u'CDR1 Seq', u'CDR1 Length', u'CDR2 Seq', u'CDR2 Length', u'CDR3 Seq', u'CDR3 Length', u'CDR3 Seq DNA', u'CDR3 Length DNA', u'Strand', u'CDR3 Found How', u'Functionality', 'V-REGION identity %', 'V-REGION identity nt', 'D-REGION reading frame', 'AA JUNCTION', 'Functionality comment', 'Sequence', 'FR1-IMGT', 'FR2-IMGT', 'FR3-IMGT', 'CDR3-IMGT', 'JUNCTION', 'J-REGION', 'FR4-IMGT', 'P3V-nt nb', 'N1-REGION-nt nb', 'P5D-nt nb', 'P3D-nt nb', 'N2-REGION-nt nb', 'P5J-nt nb', '3V-REGION trimmed-nt nb', '5D-REGION trimmed-nt nb', '3D-REGION trimmed-nt nb', '5J-REGION trimmed-nt nb'] 67 outFrame.columns = [u'ID', u'VDJ Frame', u'Top V Gene', u'Top D Gene', u'Top J Gene', u'CDR1 Seq', u'CDR1 Length', u'CDR2 Seq', u'CDR2 Length', u'CDR3 Seq', u'CDR3 Length',
68 u'CDR3 Seq DNA', u'CDR3 Length DNA', u'Strand', u'CDR3 Found How', u'Functionality', 'V-REGION identity %', 'V-REGION identity nt', 'D-REGION reading frame',
69 'AA JUNCTION', 'Functionality comment', 'Sequence', 'FR1-IMGT', 'FR2-IMGT', 'FR3-IMGT', 'CDR3-IMGT', 'JUNCTION', 'J-REGION', 'FR4-IMGT', 'P3V-nt nb',
70 'N-REGION-nt nb', 'N1-REGION-nt nb', 'P5D-nt nb', 'P3D-nt nb', 'N2-REGION-nt nb', 'P5J-nt nb', '3V-REGION trimmed-nt nb', '5D-REGION trimmed-nt nb', '3D-REGION trimmed-nt nb',
71 '5J-REGION trimmed-nt nb']
67 72
68 """ 73 """
69 IGHV[0-9]-[0-9ab]+-?[0-9]?D? 74 IGHV[0-9]-[0-9ab]+-?[0-9]?D?
70 TRBV[0-9]{1,2}-?[0-9]?-?[123]? 75 TRBV[0-9]{1,2}-?[0-9]?-?[123]?
71 IGKV[0-3]D?-[0-9]{1,2} 76 IGKV[0-3]D?-[0-9]{1,2}